Cervical cancer cell identification & detection using fuzzy C mean and K nearest neighbor techniques

Abstract

Across the globe, woman has been diagnosed two major forms of cancer, in which one is identified as cervical cancer and its micro classification. Morphology changes in cells or dead nucleus in the cervix causes cervical cancer. These cells are characterized with multiple nucleuses, faulty & lack of cytoplasm and so on. Detection of cervical cancer using smear test is extremely challenging because such cells does not offer texture variations or any significant color from the normal cells. Therefore for identification in abnormality of cells we required high level Digital image processing technique which compromises an automated, comprehensive machine learning skills. An advanced Fuzzy based technique has been implied to separate nucleus and cytoplasm from the cell. KNN is instructed with the color features and shape features of the segmented units of the cell and then an unknown cervix cell samples are classified by this technique. The proposed technique gives shape and color features of nucleus and cytoplasm of the cervix cell.